= Liquid Media's Solr/Rails Integration Library

== Setup

=== Install all your files

1. Unpack this directory somewhere useful; it's your SOLR_HOME.

2. Create a directory to hold the data; it's your SOLR_DATA. Wise people will put SOLR_DATA somewhere other than underneath SOLR_HOME or underneath RAILS_ROOT.

3. Put solr.yml and solr.rb in your RAILS_ROOT/config directory. Sample files appear below. Be sure to adjust solr.yml appropriately to reflect your file layout.

4. Put (script/)solr in your RAILS_ROOT/script directory.

These instructions are not great. I'm still working on design. To get this thing running, talk to Paul.


=== Get things going
From inside RAILS_ROOT, run the following. You should get more-or-less the same thing:

trudeau:/web/wego/wegowego paul$ script/console
Loading development environment (Rails 2.1.2)
>> @solr.ping
=> false
>> @solr.inspect
=> "#<Solr::Connection:0x23f92a4 @connection=#<Net::HTTP localhost:8982 open=false>, @autocommit=false, @url=#<URI::HTTP:0x11fc89e URL:http://localhost:8982/solr>>"
>> @solr.query('hello')
=> #<Solr::Response::Standard:0x1ff7ad4 @raw_response="{'responseHeader'=>{'status'=>0,'QTime'=>1,'params'=>{'wt'=>'ruby','q'=>'hello','fl'=>'*,score','qt'=>'standard'}},'response'=>{'numFound'=>0,'start'=>0,'maxScore'=>0.0,'docs'=>[]}}", @data={"response"=>{"maxScore"=>0.0, "docs"=>[], "start"=>0, "numFound"=>0}, "responseHeader"=>{"status"=>0, "QTime"=>1, "params"=>{"qt"=>"standard", "fl"=>"*,score", "q"=>"hello", "wt"=>"ruby"}}}, @response={"maxScore"=>0.0, "docs"=>[], "start"=>0, "numFound"=>0}, @header={"status"=>0, "QTime"=>1, "params"=>{"qt"=>"standard", "fl"=>"*,score", "q"=>"hello", "wt"=>"ruby"}}>

=== Hook up your models
I need to re-do the instructions here.. Yes, I really do.

Make sure that the solr index is updated appropriately:

  after_save :update_solr
  after_destroy :destroy_solr

  def update_solr
    @solr = Solr::Connection.new(SOLR_URL, :autocommit => :on)
    # this is a really bad duplication of @@solr_mapping, but it's all on the road to understanding the problem and possibly writing solr-ar
    @solr.update(
      :id => self.id,
      :name_t => self.name,
      :alliance_region_facet => self.metadata['alliance_region'],
      :state_facet => self.metadata['state'],
      :author_facet => self.metadata['author'],
      :year_facet => self.metadata['year'],
      :publication_type_facet => self.metadata['publication_type'],
      :subject_category_facet => self.subject_categories.map {|x| x.name}.join(', '),
      :case_name_facet => self.metadata['case_name'],
      :common_name_facet => self.metadata['common_name'],
      :text => self.metadata.to_s,
      :all_search_s => 'all'
    )
  end
  
  def destroy_solr
    @solr = Solr::Connection.new(SOLR_URL, :autocommit => :on)
    @solr.delete(self.id)
  end



== Solr.rb
require 'solr'

begin
  if File.exists?(RAILS_ROOT+'/config/solr.yml')
    config = YAML::load_file(RAILS_ROOT+'/config/solr.yml')
    SOLR_HOST = config[RAILS_ENV]['host']
    SOLR_PORT = config[RAILS_ENV]['port']
    SOLR_DATA_PATH = config[RAILS_ENV]['data_path']
    SOLR_PATH = config[RAILS_ENV]['solr_path']
    SOLR_JAR_NAME = config[RAILS_ENV]['jar_name']
    SOLR_URL = "http://#{SOLR_HOST}:#{SOLR_PORT}/solr"
  end
  @solr = Solr::Connection.new(SOLR_URL)
  @solr.ping
rescue 
  raise "Couldn't connect to the Solr server at #{SOLR_URL}. #{$!}"
  return false
end




== solr.yml
# Config file for Liquid Media Solr-Ruby integration

development:
  host: localhost
  port: 8982
  data_path: "/web/lta/solr-data/development"
  solr_path: "/web/lta/solr-clearinghouse"
  jar_name: "clearinghouse.jar"

test:
  host: localhost
  port: 8981
  data_path: "/web/lta/solr-data/test"
  solr_path: "/web/lta/solr-clearinghouse"
  jar_name: "clearinghouse.jar"

production:
  host: localhost
  port: 8984
  data_path: "/web/lta/solr-data/production"
  solr_path: "/web/lta/solr-clearinghouse"
  jar_name: "clearinghouse.jar"

== script/solr
#!/usr/bin/env ruby
require File.dirname(__FILE__) + '/../config/boot'
require 'yaml'
require 'net/http'

begin
  config = YAML::load_file(RAILS_ROOT+'/config/solr.yml')
rescue
  raise "Can't load solr.yml in #{RAILS_ROOT}/config"
end

SOLR_HOST = config[RAILS_ENV]['host']
SOLR_PORT = config[RAILS_ENV]['port']
SOLR_DATA_PATH = config[RAILS_ENV]['data_path']
SOLR_PATH = config[RAILS_ENV]['solr_path']
SOLR_JAR_NAME = config[RAILS_ENV]['jar_name']
SOLR_URL = "http://#{SOLR_HOST}:#{SOLR_PORT}/solr"

#puts "HOST: #{SOLR_HOST}; PORT: #{SOLR_PORT}; DATA_PATH: #{SOLR_DATA_PATH}; JAR_NAME: #{SOLR_JAR_NAME}"
#puts "SOLR_URL: #{SOLR_URL}"

case ARGV[0]
when 'start'
  begin
    n = Net::HTTP.new(SOLR_HOST, SOLR_PORT)
    n.request_head('/').value 
    
  rescue Net::HTTPServerException #responding
    puts "Port #{SOLR_PORT} in use" and return
    
  rescue Errno::ECONNREFUSED #not responding
    Dir.chdir(SOLR_PATH) do
      pid = fork do
        #STDERR.close
        exec "nohup java -DSTOP.PORT=#{SOLR_PORT-20} -DSTOP.KEY=ftasolrstop -Dsolr.data.dir=#{SOLR_DATA_PATH} -Djetty.port=#{SOLR_PORT} -jar #{SOLR_JAR_NAME} > #{RAILS_ROOT}/log/solr.#{RAILS_ENV}.log 2> #{RAILS_ROOT}/log/solr.#{RAILS_ENV}.err"
      end
      sleep(5)
      File.open("#{RAILS_ROOT}/tmp/pids/solr-#{ENV['RAILS_ENV']}.pid", "w"){ |f| f << pid}
      puts "Solr started successfully on #{SOLR_PORT}, pid: #{pid}."
    end
  end
when 'stop'
  Dir.chdir(SOLR_PATH) do
    file_path = "#{RAILS_ROOT}/tmp/pids/solr-#{ENV['RAILS_ENV']}.pid"
    if File.exists?(file_path)
      # This used to happen:
      # File.open(file_path, "r") do |f| 
      #   pid = f.readline
      #   Process.kill('TERM', pid.to_i)
      # end
      # But now we use Java to stop the Solr process
      fork do
        exec "/usr/bin/java -DSTOP.PORT=#{SOLR_PORT-20} -DSTOP.KEY=ftasolrstop -jar #{SOLR_JAR_NAME} --stop"
      end
      File.unlink(file_path)
      puts "Solr shutdown successfully."
    else
      puts "Solr is not running.  I haven't done anything."
    end
  end
else
  puts "Usage: script/solr <start|stop>"
end